To examine the relation between fatal prostate cancer and family history of prostate cancer in a first-degree relative, we analyzed data from a prospective mortality study of 481,011 men with no history of cancer at enrollment in 1982. During 9 years of follow-up, 1,922 deaths from prostate cancer occurred. Results from Cox proportional hazard models showed that family history of prostate cancer was related to fatal prostate cancer [rate ratio (RR) = 1.60; 95% confidence interval (CI) = 1.31-1.97]; men with two or more affected relatives had a greater than threefold increase in risk (RR = 3.19; 95% CI = 1.51-6.71). Men whose relatives were diagnosed with prostate cancer before age 65 years (RR = 2.03; 95% CI = 1.33-3.09) had a greater effect of family history than men whose relatives were diagnosed at older ages (RR = 1.50; 95% CI = 1.17-1.91). Rate ratios did not increase with decreasing age of the study participants. The 60% increase in risk for men with at least one affected relative is lower than that reported in previous studies.